The goal of this proposal is to dramatically improve PET image quality using real time motion information provided by simultaneous MR imaging in PET-MRI scanners. PET imaging of the thorax is significantly limited by poor spatial resolution due to voluntary as well as internal, i.e., cardiac and respiratory, motion. Prototype small bore simultaneous MR-PET systems now exist, and whole-body MR-PET systems are under development. These systems allow measurement of the subject's 3D non-rigid motion field using MRI without additional radiation dose. We propose to develop a novel iterative PET reconstruction framework that models the MR-derived 3D non-rigid motion field in the emission and attenuation maps. We hypothesize that the resulting PET data will have markedly improved spatial resolution, signal-to-noise ratio (SNR) and quantitative accuracy. We will evaluate the performance of the methods to be developed in realistic Monte Carlo simulations, and test their feasibility in a rabbit model for the diagnoses of small thoracic tumors. PUBLIC HEALTH RELEVANCE: The goal of this proposal is to dramatically improve the quality of Positron Emission Tomography by using the motion information available from real-time MRI in simultaneous PET-MRI scanners. This work would allow to freeze cardiac and respiratory motion without additional dose to the patient. We propose to develop novel reconstruction methods and evaluate them in numerical simulations and in animal studies that mimic small thoracic tumors.